Skipper

Here is the first graph.

An update will follow when more date is received.

The data used, has been collected in this and other forums. Data lying outside a certain deviation from the main data pool has been considered invalid, and is excluded.
When more data arrives, we will get a better, and more valid result.

However, there is so far a very clear relation between a given CPU temperature, and the amount of OC being possible.

BE AWARE!!
This is only meant as information, and you should not base any decisions on it. There is a huge amount of possible influences, that will affect your systems possibilities to reach a certain OC. You risk harming your computer if you attempt to OC it. I can and will not be held responsible for any use of this information.

Attached Images

OC.Graph.bmp (96.1 KB, 82 views)

__________________
Why do so many replies seem to come from people who are afraid of falling asleep!!!!!

Some usefull Raid and HD Info:
http://www.overclock.net/hard-drives...knowledge.html

General Ralations between CPU Temperature and OC'ing:
http://www.overclock.net/water-cooli...ons-graph.html

rippon

Wow, great work. I never suspected it to be so linear.

However I would like to comment that some of the factors affecting the % overclock are not to be ignored. Such as stepping, for example. Your graph, while a general indicator of OC/TEMP relation, could be far more informative if broken into several subcategories, most prevalent to my mind, a specific stepping of a CPU.

Attached is a .jpg version of the graph so it displays in the thread.

Attached Thumbnails

 

Skipper

Yes it is quite linear. That is partly due to the fact, that i only used 80% of the data. The discarded 20% is those lying farthest from the main data pool.

I discarded such a big percentage, due to the fact, that many data was obviously faulty, and also, to compensate for the very "enthusiastic" info from some people. Basically I weeded out the overly optimistic/pessimistic, and the downright hilarious data.

For the stepping part of it, yes that would be nice. Problem is, that I need a MUCH bigger data pool to even think about it. I barely have enough data to make this "all CPU" pool. But if there should be sufficient data for it in time, I will be happy to make it so.

What I also should explain is, that those CPU's that perform outright lousy, is weeded out of the data pool via the 80% cutoff. This means, that in order to reach close to an OC in the range of the Graph, you will have to make sure to get the right CPU. I have no insight in that, and info should be sought in other places.

__________________
Why do so many replies seem to come from people who are afraid of falling asleep!!!!!

Some usefull Raid and HD Info:
http://www.overclock.net/hard-drives...knowledge.html

General Ralations between CPU Temperature and OC'ing:
http://www.overclock.net/water-cooli...ons-graph.html

rippon

Yes, a greater range of data would be nice. I can add that I can get a 630MHz overclock, or 31.5% higher than my 2GHz stock, with load (and I mean full load with hot ambient) tempatures of 53C. Maybe not all of the data you threw out is really that far-fetched? It is hard to tell, but if you want to improve the graph, cruise over to the "max stable overclock" thread.

AMD thread link-

intel thread link-

Lots of good data in those